KR20060108648A - Methods and apparatus for optimizing a substrate in a plasma processing system - Google Patents

Abstract

An arrangement for processing a semiconductor substrate in a plasma processing system is disclosed. The arrangement includes providing a RF coupling structure having a first terminal and a second terminal, the first terminal being coupled with a first electrical measuring device, the second terminal being coupled with a second electrical measuring device. The arrangement also includes coupling a compensating circuit to the second terminal. The arrangement further includes providing a feedback circuit coupled to receive information from the first electrical measuring device and the second electrical measuring device, an output of the feedback circuit being employed to control the compensating circuit in order to keep a ratio between a first electrical value at the first terminal and a second electrical value at the second terminal substantially at a predefined value.

Description

METHODS AND APPARATUS FOR OPTIMIZING A SUBSTRATE IN A PLASMA PROCESSING SYSTEM

Background of the Invention

The present invention relates generally to substrate production techniques and specifically to methods and apparatus for optimizing substrates in a plasma processing system.

In the processing of substrates, for example glass panels or semiconductor wafers such as those used in flat panel display production, plasma is often used. As part of the substrate processing (chemical vapor deposition, plasma enhanced chemical vapor deposition, physical vapor deposition, etc.), for example, the substrate is divided into a plurality of dies, or rectangular regions, each of which will be an integrated circuit. The substrate is then processed in a series of steps where the material is selectively removed (etched) and deposited (deposited) to form an electrical element.

In an exemplary plasma process, the substrate is coated with a thin film of strengthening emulsion (ie, a photoresist mask) prior to etching. The region of the strengthening emulsion is then selectively removed and the base layer portion is exposed. The substrate is then placed in a plasma processing chamber called a chuck on a substrate support structure comprising a monopolar or bipolar electrode.

Suitable plasma processing gases (e.g., C ₄ F ₈ , C ₄ F ₆ , CHF ₃ , CH ₂ F ₃ , CF ₄ , CH ₃ F, C ₂ F ₄ , N ₂ , O ₂ , Ar, Xe, He, H ₂ , NH ₃ , SF ₆ , BCl ₃ , Cl ₂ , WF _6, etc.) are introduced into the chamber and ionized by a first RF energy source, usually coupled to an RF coupling structure, such as an induction coil set. do. In addition, a second RF energy source may also be coupled to the substrate to create a bias with the plasma and direct the plasma away from the plasma processing system to the substrate.

Induction coils are devices similar in purpose to transformers, which allow time-varying voltages and potential differences in the plasma processing gas to generate plasma by the current being turned on and off sequentially in the base coil. A common configuration is to place a planar coil on top of the chamber, also called TCP ^TM (ie Transformer coupled plasma). Another configuration constitutes a plasma processing system such that the solenoid coil is wound around the side of the plasma processing chamber.

In general, plasma consists of partially ionized gases. Since the plasma discharge is RF driven and weakly ionized, the electrons in the plasma are not in thermal equilibrium with the ions. That is, while heavy ions efficiently exchange energy by collision with a background gas (eg argon, etc.), electrons absorb thermal energy. Since the electrons have a substantially smaller mass than the ions, the electron heat rate is much greater than the ion heat rate. This allows faster moving electrons to deviate from the surface in the plasma processing system, creating a substantially positively charged ion envelope between the plasma and the surface. The ions entering the envelope are then accelerated to the surface.

Low RF frequencies allow plasma ions to pass through the shell in less than one RF cycle, producing a large change in ion energy. In addition, high RF frequencies allow plasma ions to pass through the shell through several RF cycles, creating a more consistent set of ion energies. Higher frequencies cause lower envelope voltages than when lifted by low frequency signals at similar power levels.

The coupling between the RF source and the plasma processing chamber is generally a matching network. In general, from a transmission line termination point of view, the matching network converts the complex impedance of the plasma to the nominal output impedance of the RF generator. For example, if an RF generator delivers 2 kW of output power (called a bag or transmit power) and the matching network is not properly "tuned" (eg, causes 50% reflected power), then 1 kW of power. It will reflect back to this RF generator. This means that only 1 kW is delivered to the rod (plasma chamber). The combination of high quality, low impedance, suitably selected length transmission lines with matching networks of appropriate size (for current and impedance ranges) can provide the best power transfer from the generator to the plasma chamber.

Referring now to FIG. 1, a simplified diagram of a plasma processing system 100 is shown. In general, a set of suitable gases enters the chamber 102 from the gas dispersion system 122 through the inlet 108. In order to process (eg, etch or deposit) an exposed area of the substrate 114, such as a semiconductor wafer or glass plate, located on the electrostatic chuck 116, the plasma processing gases are substantially ionized to form a plasma 110. To form. Together with the liner 112, the upper chamber plate 120 allows the plasma 110 to be optimally focused on the substrate 114.

Gas dispersion system 122 is commonly used for plasma processing gases (eg, C ₄ F ₈ , C ₄ F ₆ , CHF ₃ , CH ₂ F ₃ , CF ₄ , HBr, CH ₃ F, C ₂ F ₄ , N ₂ , O ₂ , Ar, Xe, He, H ₂ , NH ₃ , SF ₆ , BCl ₃ , Cl ₂ , WF _6, etc.). Gas cylinders 124a-f may also be protected by enclosure 128 providing local exhaust ventilation. Mass flow controllers 126a-f are commonly used in the semiconductor industry to include a complete device (including transducers, control valves, and control and signal processing electronics) to measure and control the gas mass flow to the plasma processing system. to be.

Induction coil 131 is isolated from the plasma by dielectric window 104 and generally induces a time varying electrical current in the plasma processing gas to produce plasma 110. The window protects the induction coil from the plasma 110 and allows the generated RF field to penetrate into the plasma processing chamber. Also coupled to induction coil 131 in leads 130a-b is matching network 132, which may be coupled to RF generator 138. As mentioned above, matching network 132 attempts to match the impedance of the RF generator with the impedance of plasma 110, which typically operates at 13.56 MHz and 50 ohms.

Referring now to FIG. 2, a simplified diagram of a TCP ^™ induction coil is shown. Induction coils may be manufactured from highly conductive copper tubes-generally round, square or square, depending on the application and must be robust to withstand certain uses. As shown in FIG. 1, the leads 130a-b are used to couple the induction coil 131 to the matching network 132.

However, as the chamber pressure or power level changes, the matching network and load may become unstable together. The result is a sudden change or jitter that changes faster than the matching network can respond to. The resulting power transfer instability not only damages the components of the matching network and the RF generator, but can also substantially affect the productivity of the substrate.

In view of the foregoing, a method and apparatus for optimizing a substrate in a plasma processing system is desired.

Summary of the Invention

In one embodiment, the invention is directed to an array method for processing a semiconductor substrate in a plasma processing system. The method includes providing an RF coupling structure having a first terminal and a second terminal, wherein the first terminal is coupled with the first electrical measurement device, and the second terminal is coupled with the second electrical measurement device. There is. The method also includes coupling a compensation circuit to the second terminal. The method also includes providing a feedback circuit coupled to receive information from the first electrical measurement device and the second electrical measurement device, the first electrical value at the first terminal and the second at the second terminal. The output of the feedback circuit is used to control the compensation circuit to maintain the ratio between the electrical values to a substantially predetermined value.

In another embodiment, the present invention is directed to an apparatus for processing a semiconductor substrate in a plasma processing system. The apparatus includes means for providing an RF coupling structure having a first terminal and a second terminal, wherein the first terminal is coupled with the first electrical measurement device and the second terminal is coupled with the second electrical measurement device. The device also includes means for coupling the compensation circuit to the second terminal. The apparatus also includes means for providing a feedback circuit coupled to receive information from the first electrical measurement device and the second electrical measurement device, the first electrical value at the first terminal and the second at the second terminal. The output of the feedback circuit is used to control the compensation circuit to maintain the ratio between the electrical values to a substantially predetermined value.

These and other features of the present invention will be described in more detail in conjunction with the following detailed description of the invention and the following figures.

Brief description of the drawings

The invention is illustrated by way of example and not by way of limitation in the figures of the accompanying drawings in which like reference numerals designate like elements.

1 shows a simplified diagram of a plasma processing system.

2 shows a simplified diagram of TCP ^™ .

3 shows a simplified diagram of a matching network, in accordance with an embodiment of the present invention.

4 shows a simplified model of a plasma rod 310, in accordance with an embodiment of the present invention.

5 illustrates a simplified electrical diagram of a matching network in which the induction coil voltage is substantially optimized and stabilized, in accordance with an embodiment of the present invention.

FIG. 6 illustrates a portion of the matching network of FIG. 5 in which the voltage amplitude is stable, in accordance with an embodiment of the present invention.

Detailed Description of the Preferred Embodiments

The invention will be described in detail with reference to several preferred embodiments as described in the accompanying drawings. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. However, it will be apparent to one skilled in the art that the present invention may be practiced without some or all of these specific details. In other instances, well known process steps and / or structures have not been described in detail in order not to unnecessarily obscure the present invention.

Without wishing to be bound by theory, it is believed by the inventors that voltage balance across an RF coupling structure, such as an induction coil, may optimize the plasma in a plasma processing system. For example, the plasma may be generated by an inductively coupled azimuthal electric field produced by the induction coil. Basically in a planar antenna, the induction coil is often a square coil spiral coil located on top of the dielectric window. The voltage applied to the antenna at 13.56 MHz creates a vibrating magnetic field around the coil, which penetrates the plasma and produces a counter electric field.

In an ideal plasma processing system, the azimuth electric field is zero in axial and zero in plane, thereby having a peak value in an annular region that is approximately half the radius. In a carefully designed plasma processing system, the ratio of capacitively coupled power from the coil to the plasma can be small, thereby creating a plasma potential that does not vibrate vividly during radio frequency cycles. As mentioned above, ion energy can be controlled by independently applying RF potential and create a bias in the substrate.

However, since the plasma load of the matching network may change when the substrate is processed and the plasma is consumed, and the matching network may be optimized for the expected plasma load in the fabrication of the plasma processing system, the induction coil may be removed during processing of the substrate. The actual voltage balance that passes may not be optimal for a wide range of process conditions or changing conditions. The resulting electric field may be radially deformed, which may result in non-uniform plasma density substantially across the substrate, potentially affecting yield.

This condition becomes even more problematic as the demand for high density circuits continues to increase. For example, in a plasma etching process, faceting may occur if the plasma is not properly optimized. Facets are nonlinear contours on a substrate, such as trench sides. Areas of low density plasma may not remove a sufficient amount of material from the substrate and substantially reduce the size of the trench or via. In addition, areas of high density plasma may remove excess material from the substrate and create a substantially undercut of the cavity.

Referring to FIG. 3, a simplified diagram of a matching network is shown in accordance with one embodiment of the present invention. In general, measuring electrical characteristics, such as voltage, through an electrical measurement device coupled to an induction coil, such as a voltage probe, may not provide sufficient information to estimate the power or impedance delivered. Indeed, in an unclear way, a model of a match network has been developed to anticipate the load impedance from the match network setting. Since the voltage on the coil is known, the delivered power may be calculated from the estimated load impedance. Estimated in this way, the value of power has an error of about 9%.

The detailed model of the matching network 332 can include not only visible elements, such as vacuum capacitors, but also inductances of all connections between components. The current flowing through this connection leads to other currents, especially in the walls of the match network enclosure. This induced current forms part of the complex current back to the RF generator. Resistance loss in the wall contributes to the effective resistance of the match network.

However, this model is very complex and can contain a large number of elements, making each quantification difficult. A simpler approach models a matching network that includes only a set of lumped components. For example, inductor 331 corresponds to the connection between variable capacitors C1 and C3, which may be about 15 inches long. It is assumed to have an inductance of about 100 nH. The remaining connections are only a few inches long and are ignored. All elements other than C3 are considered part of the plasma rod 310 (e.g., induction coil, plasma, substrate, etc.). Leads 303a-b connecting the ends of the TCP ^™ coils together are about 12 inches and are part of the load impedance.

Each vacuum capacitor may have a unique series inductance of about 20 nH, but these are not explicitly included. This series inductance is not detected when the capacitors are calibrated at a frequency of 1 kHz and only barely identifiable at a starting frequency of 13.56 MHz. Current induced in the network housing reduces the effective inductance of the component causing induction. This can be considered by the empirical choice of inductance that best fits the data.

 The resistance of the match network is assumed to be small so that the conjugate impedance method can be used to calculate the load impedance from the capacitor setting. Capacitors C1 and C3 are calibrated against the count and a typical calibration curve is shown, from which capacity can be easily determined. The values may be between about 28 pF and 480 pF in this particular network. Capacitors C2 and C4 are fixed at 102 and 80 pF, the standard values corresponding to a 2000 mm STAR configuration.

V3 and V4 are electrical measurement devices, such as RF voltage probes, located near each end of an induction coil that alternately supplies induced RF power to the plasma chamber. The electrical measuring device V3 may measure a first voltage between the matching network and the induction coil, and the electrical measuring device V4 measures the second voltage between the load and the termination capacitor C4. In one embodiment, the electrical measurement device is similar in design to the RF voltage peak detector. In general, voltage probes allow RF voltage by 5 kV (peak) and are measured without an arc or other discharge. In general, the absolute accuracy of the probes is about 3% from about 200V to about 5kV.

Referring to FIG. 4, a simplified model of a coil and plasma rod 310 is shown, in accordance with an embodiment of the present invention. In general, the coil and plasma rod 310 include a plasma processing system element that is electrically positioned between V3 and V4 (eg, induction coil, plasma, substrate, etc.). Like the matching network, the coil and plasma rod 310 may be modeled as a transformer circuit comprising a parallel resistor 408 that provides core power loss of the plasma, and a parallel inductance 404, which translates the magnetic flux into the plasma. Explain the bonding.

Referring to FIG. 5, in accordance with one embodiment of the present invention, a simplified electrical diagram of a matching network in which the induction coil voltage is substantially optimized and stabilized is shown. In a non-obvious manner, in order to adjust the ratio of the first voltage to the second voltage, a feedback circuit is coupled between compensation circuit C5, such as a variable resistor, between V3 and V4.

According to FIG. 3, a simplified matching network model as a set of lumped components, in accordance with an embodiment of the present invention, is shown. Inductor 331 corresponds to the connection between C1 and C3. All elements other than C3 are considered part of the coil and plasma rod 310 (eg, induction coil, plasma, substrate, etc.). Connected to the end of the TCP ^TM coil Leads 303a-b are part of the load impedance. V3 and V4 are located near each end of the induction coil which in turn supplies inductive RF power to the plasma chamber. V3 measures the first voltage between the matching network and the induction coil, and V4 measures the second voltage between the load and the termination capacitor C5.

The feedback circuit 502 is a circuit that includes a forward path 504, a signal path including a feedback path 506, and is coupled to a capacitor C5. In a non-obvious manner, when the substrate is processed and the plasma is consumed, the feedback circuit 502 optimizes the induction coil voltage in a substantially dynamic manner. In one embodiment, C5 is automated. In other embodiments, the feedback circuit 502 may then adjust C5 during substrate processing to maintain the first voltage and the second voltage substantially stable. In another embodiment, the feedback circuit can adjust the ratio of the first voltage to the second voltage to any desired value K to provide plasma processing gain, such as etch uniformity tuning. In another aspect of the invention, the preferred range of ratio (K) is between about 0.5 and about 1.5. In another aspect of the invention, the more preferred range of ratio K is between about 0.75 and about 1.25. In another aspect of the invention, the most preferred range of ratio (K) is about 1.

With reference to FIG. 6, shown in FIG. 5 is a portion of the matching network of FIG. 5 in which voltage amplification is balanced. Without wishing to be bound by theory, it is believed here by the inventor that the optimum balance configuration is 180 ° with respect to the first voltage. In general, when the voltage phases of the two RF powers are approximately equal, the plasma spreads and the density and process speed decrease. However, when the voltage phase difference is about 180 degrees, the plasma density becomes substantially higher.

As mentioned above, from the viewpoint of the transmission line termination, the matching network 332 converts the complex impedance of the plasma to a frequency of about stable 50 ohms and about 13.56 MHz. To ensure that the coil and plasma rod 310 are optimized, the feedback circuit 502 monitors the voltage at V3 and V4 via the feedback path 506, adjusts C5, and the voltage value at V3 is substantially V4. It is equal to the voltage value in, and has a phase 602 that is about 180 ° with respect to the phase 604. Also in another embodiment, the feedback circuit 502 is coupled to the diagnostic monitoring device. For example, if the measured voltage or impedance value falls outside a certain process range, a warning message may be sent to the preventive maintenance system for further investigation.

Although the invention has been described in several preferred embodiments, there are variations, substitutions, and equivalents falling within the scope of the invention. For example, the present invention may be used in a ram research plasma processing system (e.g., Exelan ^™ , Exelan ^™ HP, Exelan ^™ Although described in conjunction with HPT, 2300 ^™ , Versys ^™ Star, etc., other plasma processing systems may be used. The invention may also be used as substrates of various diameters (eg, 200 mm, 300 mm, etc.). In addition, other electrical components, such as a variable inductor, may be used in place of the variable capacitor. In addition, the electrical measurement device described herein may measure electrical characteristics such as current, impedance, etc., but not voltage. There are also many ways of implementing the methods of the present invention.

Advantages of the present invention include methods and apparatus for optimizing a substrate in a plasma processing system. Additional advantages include optimizing plasma density throughout the plasma process and providing diagnostic information for preventive maintenance.

In addition to the disclosed exemplary embodiments and best modes, modifications and variations may be made to the disclosed embodiments that fall within the spirit and scope of the invention as defined by the following claims.

Claims (32)

An arrangement method for processing a semiconductor substrate in a plasma processing system, comprising Providing an RF coupling structure having a first end and a second end, the first end coupled to a first electrical measurement device, and the second end coupled to a second electrical measurement device; Coupling a compensation circuit to the second end; And Providing a feedback circuit coupled to receive information from the first electrical measurement device and the second electrical measurement device, the output of the feedback circuit being a first electrical value at the first end and the second end; Used to control the compensation circuit to maintain a ratio between the second electrical value at substantially a predetermined value, And an arrangement method for processing the semiconductor substrate. The method of claim 1, And the first electrical value is a voltage. The method of claim 1, And the first electrical value is a current. The method of claim 1, And the second electrical value is a voltage. The method of claim 1, And the second electrical value is a current. The method of claim 1, And the second compensation circuit is a variable capacitor. The method of claim 1, And the second compensation circuit is a variable inductor. The method of claim 1, And the second compensation circuit and the feedback circuit are integrated. The method of claim 1, And the second compensation circuit is coupled to ground. The method of claim 1, And the first end is coupled to a matching network. The method of claim 1, And the matching network is coupled to an RF generator. The method of claim 1, And the predetermined value is between about 0.5 and about 1.5. The method of claim 1, And the predetermined value is between about 0.75 and about 1.25. The method of claim 1, And the predetermined value is about one. The method of claim 1, And the feedback circuit is microprocessor controlled. The method of claim 1, Wherein the predetermined value may be set to a new value for each step of the process method. An apparatus for processing a semiconductor substrate in a plasma processing system, Means for providing an RF coupling structure having a first end and a second end, the first end coupled with a first electrical measurement device and the second end coupled with a second electrical measurement device; Means for coupling a compensation circuit to the second end; And Means for providing a feedback circuit coupled to receive information from the first electrical measurement device and the second electrical measurement device, the output of the feedback circuit being a first electrical value at the first end and the second end; Means for controlling the compensation circuit to maintain a ratio between the second electrical value at substantially a predetermined value And an apparatus for processing a semiconductor substrate. The method of claim 17, And the first electrical value is a voltage. The method of claim 17, And the first electrical value is a current. The method of claim 17, And the second electrical value is a voltage. The method of claim 17, And the second electrical value is a current. The method of claim 17, And the second compensation circuit is a variable capacitor. The method of claim 17, And the second compensation circuit is a variable inductor. The method of claim 17, And the second compensation circuit and the feedback circuit are integrated. The method of claim 17, And the second compensation circuit is coupled to ground. The method of claim 17, And the first end is coupled to a matching network. The method of claim 17, And the matching network is coupled to an RF generator. The method of claim 17, And the predetermined value is between about 0.5 and about 0.15. The method of claim 17, And the predetermined value is between about 0.75 and about 1.25. The method of claim 17, And the predetermined value is about one. The method of claim 17, And the feedback circuit is microprocessor controlled. The method of claim 17, And the predetermined value may be set to a new value at each step of the process method.

Images